Duffy Status and Hydroxyurea Dosing in Adult Patients with Sickle Cell Disease: A Potential for Iatrogenic Harm?

Introduction:

A lower absolute neutrophil count (ANC) is associated with the variant rs2814778 (c.1-67T>C) in the promotor of the ACKR1 gene. Homozygosity for rs2814778 results in the Duffy null phenotype. Duffy null patients have normal bone marrow cellularity and are not at increased risk of infection. Duffy null individuals have malaria survival advantage as the Duffy antigen is the erythrocyte receptor for Plasmodium Vivax, which probably accounts for the allele frequency of 0.81 in African Americans (Afr Am) compared to 0.004 in Caucasians.

Among Afr Am, 68% are Duffy null, 23.8% of whom have ANC of <2 k cells/mL and 10% have ANC <1k cells/mL. Sickle cell disease (SCD), a complex multi-systemic disorder, affects ~7-8 million people worldwide, predominantly of African descent and 1 in 365 Afr Am births. Hydroxyurea (HU), a myelosuppressive agent, is standard care in SCD. As HU dosing in SCD is based on achieving maximum tolerated dose (MTD) typically targeting ANC of 1.5-3 x 10³ /uL, there is concern about potential underdosing and iatrogenic harm in patients who are Duffy-null.

Previous studies on HU dosing in SCD based on Duffy status have focused on children. Here, we investigate dosing differences in adult SCD patients and whether they are at risk of lower ANCs.

Method

The cohort comprised 673 adult patients with SCD enrolled under protocols NCT00011648 and NCT00081523 at the National Heart, Lung and Blood Institute (NHLBI) between Sept 2006 and Feb 2017. Demographics, HU treatment (dosing based on MTD) status, hematological indices and biochemistry on the day of sampling were recorded in a database with unique identifiers. ACKR1 rs2814778 genotype was derived from whole genome sequence data. Duffy phenotype was established by presence and zygosity of ACKR1 rs2814778. The cohort was stratified by Duffy phenotypic status as well as on/off HU therapy to evaluate differences in ANC, fetal hemoglobin (%HbF), and HU dosing (mg/kg). The comparison was conducted based on Mann-Whitney test using R (v4.3.2).

Results:

Mean age of the cohort (N=673) was 34.2 (± 12.05 years), 352 female and 321 male. Ethnicity was available for 653 (97.03%) patients of which 630 (96.48%) were Black Afr Am: 522 (82.8%) have sickle cell anemia (SCA) i.e. HbSS or HbSB0 and the remainder have non-SCA SCD (HbSC, HbSB+, HbSD or HbSOArab).

Of the 630 Afr Am subjects, 74% were homozygous for ACKR1 rs2814778 and displayed a Duffy null phenotype with an allele frequency of 0.868. The ANC was significantly higher for SCD patients with Duffy non-null compared to those with Duffy null (mean±SD: 6.33±3.17 x 10³ /mL vs 5.66±3.05; P=0.011). The elevated ANC in the Duffy non-null subjects sustained when stratifying patients based on HBB genotype (6.55±3.24 vs 5.92±3.19 in SCA patients, P=0.037; 5.41±2.75 vs 4.33±1.67 in non-SCA patients, P=0.041). Only 1% of the SCA patients and 1.23% of the non-SCA SCD patients had ANC of <1.5. There was no statistically significant difference in other markers of disease severity (Platelets, Reticulocytes, ESR, LDH) and no mortality difference between Duffy null and non-null phenotypes in SCA and non-SCA SCD subjects.

In the SCA cohort 221 (76.9% Duffy null) were on and 300 (76 % Duffy null) were off HU therapy. Among the 221 on HU, there was no statistically significant difference in ANC, mean %HbF and HU dose between Duffy null and non-null groups (ANC: 5.31±2.77 vs 5.78±2.88, P=0.26; %HbF: 10.4±6.69 vs 10.5±6.23, P=0.74; HU dose: 17.9±8.02 vs 17.1±8.68, P=0.62).

Discussion:

We confirmed that the allele frequency of ACKR1 rs2814778 in our SCD cohort is comparable to that in the general Afr Am population with a similar incidence of Duffy null phenotype. Duffy null status confers a lower ANC compared with those who express Duffy antigen, which is most marked in the non-SCA SCD patients. However, the difference is less marked when compared with published studies of healthy black patients. This may be in keeping with the higher ANC associated with the inflammatory milieu of SCD.

There is no significant difference in HU dose nor %HbF in patients stratified by Duffy status. The proportion of patients who are on HU is comparable between the Duffy null and Duffy non-null.

This important negative finding in adults is comparable to published data on pediatric patients and indicates that Duffy null phenotype does not negatively impact HU dosing.